Hydrological droughts, when combined with mismanagement of reservoir water resources, can lead to severe shortages. Reliable early warning of hydrological droughts can support improved monthly operation of upstream reservoirs. In this study, we evaluated an early-warning approach that integrates local rainfall with modeled soil moisture from the DREAM rainfall–runoff model, applied to the Camastra river basin upstream of the Camastra dam in Basilicata, Southern Italy. The performance of the proposed approach was tested against streamflow data derived by an inverse reservoir water balance.
Monthly soil moisture data were averaged over 1-, 3-, and 6-month periods and standardized per calendar month using a beta cumulative distribution function. When the Shapiro–Wilk test did not reject normality (p>0.05), the data were further transformed into Z-scores via probit mapping, producing a series of Soil Moisture Index (SMI) values. A Joint Drought Index (JDI) was then computed by combining the Standardized Precipitation Index at a 6-month aggregation time scale (SPI-6) and SMI-k (k = 1, 3, 6) through a bivariate Gaussian copula. The predictive skill of JDI was evaluated against hydrological drought events defined by Standardized Streamflow Index (SFI) thresholds (≤0, ≤–1, ≤–1.5, ≤–2) using ROC/AUC with 12-month block-bootstrap 95% confidence intervals, testing lead times at k = 0–3 months and benchmarking against SPI-6 alone.
Cross-correlation analysis revealed a coherent propagation chain: SPI-6 → SMI-6 peaked near a 1-month lead (r ≈ 0.85–0.87), while SMI-6 → SFI peaked at a 0-month lead (r ≈ 0.75). Consistently, JDI-6 showed predictive skill at lead times of 0 and 1 month. Compared to SPI-6 alone, JDI generally matched or slightly outperformed it, particularly at k = 1, demonstrating the added value of soil moisture information.
These results indicate that JDI can provide reliable short-term (0–1 month) hydrological drought warnings, particularly when direct inflow data are unavailable, thereby supporting timely operational decisions for water-supply management.
